Embed Size (px)
Citation preview
General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
Users may download and print one copy of any publication from the public portal for the purpose of private study or research.
You may not further distribute the material or use it for any profit-making activity or commercial gain
You may freely distribute the URL identifying the publication in the public portal If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from orbit.dtu.dk on: Nov 03, 2020
Profitability of Frequency Regulation by Electric Vehicles in Denmark and JapanConsidering Battery Degradation Costs
Calearo, Lisa; Marinelli, Mattia
Published in:World Electric Vehicle Journal
Link to article, DOI:10.3390/wevj11030048
Publication date:2020
Document VersionPublisher's PDF, also known as Version of record
Link back to DTU Orbit
Citation (APA):Calearo, L., & Marinelli, M. (2020). Profitability of Frequency Regulation by Electric Vehicles in Denmark andJapan Considering Battery Degradation Costs. World Electric Vehicle Journal, 11(3), [48].https://doi.org/10.3390/wevj11030048
Article
Profitability of Frequency Regulation by ElectricVehicles in Denmark and Japan Considering BatteryDegradation Costs
Lisa Calearo * and Mattia Marinelli
Center for Electric Power and Energy, Department of Electrical Engineering, Technical University of Denmark,Risø Campus, 4000 Roskilde, Denmark; [email protected]* Correspondence: [email protected]
Received: 15 May 2020; Accepted: 14 July 2020; Published: 16 July 2020�����������������
Abstract: This paper determines the profitability of the primary frequency regulation (FR) serviceconsidering the wear of the electric vehicle (EV) battery as a cost. To evaluate the profitability of theFR service, the cost of degradation from FR provision is separated from the degradation caused bydriving usage. During FR, the power response is proportional to the frequency deviation with fullactivation power of 9.2 kW, when deviations are larger than 100 mHz. The degradation due to FRis found to be an additional 1–2% to the 7–12% capacity reduction of a 40 kWh Lithium-ion NMCbattery pack over 5 years. The overall economic framework is applied in Denmark, both DK1 andDK2, and Japan, by considering historical frequencies. The DK2 FR market framework is taken asreference also for the Japanese and the DK1 cases. Electricity prices and charger efficiency are the twomain parameters that affect the profitability of the service. Indeed, with domestic prices there is noprofitability, whereas with industrial prices, despite differences between the frequencies, the serviceis similarly profitable with approx. 3500¤ for a five-year period.
Keywords: degradation; electric vehicle; frequency regulation; vehicle-to-grid
1. Introduction
Primary frequency regulation (FR) service has been demonstrated to be technically feasible andeconomically profitable in different studies worldwide [1–6]. In the Nordic grid the bi-directionality ofthe±10 kW DC chargers can give a profit 17 times higher than the unidirectional case [7,8]. The revenuefrom FR is function of different parameters: electricity, regulation market prices, plugin hours,power capacity of the charger etc. [9]. In the Great Britain depending on the usage FR profit variesbetween 60 and 400 ¤/year [10]. In France EVs are found to have a revenue of approximately100 ¤/year [11], in Germany the average revenue is 200 ¤/year [12]. Most of the analyses, particularlyin the past, evaluated the profit from FR service as difference between the revenue and the chargerlosses, without giving the proper attention to the wear the EV battery is exposed to. That is because thebattery degradation is a physicochemical process with many uncertainties, non-linearities, and differentpeculiarities depending on the chemistry considered. Therefore, to account for influence of degradation,different types of models have been derived in a semi-empirical way. Three of the most discussedmodels in the literature are: NREL model based on NCA and LFP chemistries [13], Wang modelbased on LMO-NMC chemistry [14] and MOBICUS model based on LMO/NMC chemistry [15].Therefore, in our battery implementation the Wang model was implemented, as the most completeand appropriate in terms of chemistry. Nevertheless, being the empirical models affected by differentlimitations due to time resolution, data, cycling definition and chemistry [3], the authors recommendreaders to consider also new approaches for future investigations.
World Electric Vehicle Journal 2020, 11, 48; doi:10.3390/wevj11030048 www.mdpi.com/journal/wevj
World Electric Vehicle Journal 2020, 11, 48 2 of 15
In this paper, the authors aim at evaluating the profit of the service including the batterydegradation (BD) loss. This loss is technically quantified considering a Simulink model previouslydeveloped by the authors [16] to determine the economic loss. The profit from FR service is evaluatedas difference between the revenue from FR and the sum of the costs of grid energy exchanged andBD. This paper is organized as follows. Section 2 describes the methodology for the profit evaluation,with details on the revenue, costs of grid energy exchanged and BD loss. Section 3 presents the testedstudy case, and Section 4 provides the results of the different scenarios analyzed. Section 5 presents asensitivity analysis and Section 6 concludes the article with the main outcomes. Figure 1 provides theprofit evaluation flowchart, from the technical (orange) to the economic (green) characterization whichare described in Sections 3 and 2, respectively. Pbatt is the battery power output.
Figure 1. Profit evaluation flowchart: orange for the technical analysis provided in Section 3, green forthe economic analysis provided in Section 2.
2. Profit Evaluation
In this section, the methodology to evaluate the profit of the EV user is described in detailed.By definition, the profit is the difference between the revenue and the costs, therefore the profit fromproviding FR is determined as in (1):
Pro f itFR = RevenueFR − CostFRLoss − CostFR
BD (1)
where RevenueFR is the revenue from the FR service, CostFRLoss is the difference between the cost of the
energy sold and the energy purchased with the grid, considering the charger loss, and CostFRBD is the
cost of BD due to the FR service. The next Sections present how revenue and costs are evaluated.
2.1. Revenue Assessment
In Denmark FR is a service which is paid per availability. This means that the price is based on thepower capacity that is available for an hour (MW per h) and not on the actual energy that is provided.Thus, when providing FR service, the capacity provided (Pcap) is known and the revenue is quantifiedas in (2):
RevenueFR =tFR
∑0
Pcap ∗ CP (2)
where tFR is the number of hours of service availability and CP is the hourly capacity payment.
2.2. Energy Cost Assessment
During FR service, the power provided by the battery is derived from the frequency as follows [17]:
World Electric Vehicle Journal 2020, 11, 48 3 of 15
Pbatt =
Pcap − c, if ft < 49.9 Hz
50− ft0.1 ∗ Pcap − c, if 49.9 Hz ≤ ft ≤ 50.1 Hz−Pcap − c, if ft > 50.1 Hz
(3)
where c is the offset, needed to avoid that the vehicle gets too overcharged, Pcap is the power capacityof the charger and ft is the measured frequency at time t. In this analysis Pbatt is negative when thebattery is charging, and positive when it is discharging. From the grid perspective, the power (Pgrid)provided is derived from the battery power and the charger efficiency (4):
Pgrid =
Pbatt ∗ ηd if Pbatt ≥ 0Pbattηc
if Pbatt < 0(4)
where ηd and ηc are the discharging and charging efficiencies, which depend on the different poweroutput, see Table 1. To evaluate the cost due to the energy exchanged with the grid, the energy fordriving should be subtracted from the energy consumption as in (5):
CostFRLoss = EdischFR ∗ PElsale
− EchFR ∗ PElpurch(5)
EchFR is the energy from the grid (charging the EV) and EdischFR is the energy to the grid (discharging theEV) during the FR service. PElpurch
is the purchase electricity price and PElsaleis the sale electricity price.
2.3. Degradation Cost Assessment
The battery degradation consists of capacity fade and increase of the internal resistance. For thepurpose of this article, only the capacity fade is considered, nevertheless it is in the authors’ interest toinclude the increase of internal resistance in future investigations [18].
The battery degradation is the sum of two effects: calendar aging and cycle degradation. In thismanuscript the authors considered the battery degradation of the Wang model, which is based ona large set of testing data of 1.5 Ah 18650 LMO-NMC Sanyo technology [14]. The formulation wascharacterized and validated for the considered cells, nevertheless the authors are currently working onthe validation of the formulation with the similar cells used in this manuscript [19].
In the Wang model, the calendar loss equation is derived from a combination of the Arrheniusequation and the fit of the model parameters, and then the cycle degradation is calculated as differencebetween the measurements of the total capacity loss and the calendar ones.
The calendar loss (Q∗cal), when temperature and SOC are constant, is estimated through the modelusing the Arrhenius equation in (6) [14,20]:
Q∗cal = f ∗ e−EaRT ∗ t0.5 (6)
where Q∗cal is the percentage of capacity loss induced by calendar aging, f is the pre-exponential factor,a non-linear function of both SOC and temperature [16], equal to 14,876 day−1/2 for specific SOCand temperature, Ea is the activation energy equal to 24.5 kJ mol−1. R is the gas constant equal to8.314 J mol−1 K−1. The calendar loss is function of the absolute temperature (T), temperature increasesresult on larger losses.
The percentage cycle degradation is estimated as in (7) [14,20]:
Q∗cycle = B1 ∗ eB2∗rate ∗ eqCycles (7)
World Electric Vehicle Journal 2020, 11, 48 4 of 15
B1 = a ∗ T2 + b ∗ T + c (8)
B2 = d ∗ T + e (9)
rate =|I|
Cbatt(10)
eqCycles =rate
2 ∗ 3600(11)
where a, b, c, d and e are 8.58 ∗ 10−6 Ah−1 K−2, −0.0051 Ah−1 K−1, 0.759 Ah−1, −0.0067 K−1 − (C−rate) and 2.35 (C− rate)−1, respectively. I is the current flowing inside the battery pack and Cbatt isthe initial capacity of the battery, expressed in Ah, considered in the investigation. As for the calendarloss, the cycle degradation is also function of the battery temperature, with the relation provided inFigure 2.
0 10 20 30 40 500
1.5
3
4.5
6
7.5
9·10−3
T[oC]
Q∗ cy
cle/
eqC
ycle
s[%]
Figure 2. Cycle degradation dependency on battery temperature.
It is relevant to observe that being a semi-empirical model, it does not have absolute validity,as it is limited by time resolution, data limitation, parameters definition and chemistry of the cells.For further information regarding the battery model, please refer to the authors’ works [16,21].
2.3.1. General Assessment
For EV uses the battery end-of-life is usually defined by a 20% reduction of the initial capacity.Nevertheless, a Li-ion battery can still be used when the state-of-health (SOH) is equal to 80% forsecond-hand uses. The end-of-life of second-hand batteries is set at approx. 50% SOH as shown inFigure 3.
Figure 3. Battery end-of-life evaluation.
Thus, considering the 50% capacity as usable capacity of the battery, the cost of the BD is evaluatedas in (12):
CostBD =CBD50%
∗ PricebattkWh(12)
where CBD is the capacity loss due to degradation and PricebattkWhis the price of the battery per kWh.
World Electric Vehicle Journal 2020, 11, 48 5 of 15
2.3.2. Frequency Regulation Assessment
For the evaluation of the BD costs caused by the FR service provision, two scenarios are defined:
• driving (driv): the EV is only driven• driving plus FR (driv + FR): the EV is driven and used for providing FR service.
The lost capacity in the two scenarios is evaluated as in (13):
CdrivBD = Cbatt − Cdriv
batt f
Cdriv+FRBD = Cbatt − Cdriv+FR
batt f
(13)
where Cbatt is the initial battery capacity and Cdrivbatt f
, Cdriv+FRbatt f
are the battery capacities in the two
scenarios, at the end of the simulation.Considering the 50% capacity as the full capacity of the battery, the cost of the BD in the two
scenarios is evaluated as in (14):
CostdrivBD =
CdrivBD
50%∗ PricebattkWh
Costdriv+FRBD =
Cdriv+FRBD50%
∗ PricebattkWh
(14)
It is then possible to determine the costs due to FR provision as in (15):
CostFRBD = Costdriv+FR
BD − CostdrivBD (15)
3. Test Case
3.1. Technical Characterization Battery
A 40 kWh battery composed of Lithium-ion Nickel Manganese Cobalt pouch cells is taken inconsideration in the analysis [16]. The Simulink model simulates the battery EV usage given as inputthe outside temperature and the battery power profile. The outputs of the model are: current andvoltage of the battery, temperature, state-of-charge (SOC) and degradation losses.
3.1.1. Input: Power
In the driv scenario the EV is considered to be driven 40 km, 20 km in the morning and 20 in theafternoon, and it can charge from 6:00 to 8:00 (for how long, it depends on the driving, approximately40 min), until when it reaches the requested SOC [22]. In the driving + FR scenario the EV is driving40 km and performing FR service from 17:15 to 6:00 of the day after (approx. 13 h) and then it cancharge from 6:00 to 8:00 to the required SOC. Since the battery degradation is function of the SOC,two simulations are performed: in the first one the average SOC of the simulation is equal to 55% andin the second one it is equal to 75%. To keep the average SOC at these levels, avoiding that the batterygets too charged or discharged, the model is constrained between a maximum and a minimum SOClimits of 20 and 65% in the case with 55% average SOC, and 30 and 85% in the second case.
The battery power during the FR service provision is derived from the frequency values,considering a charger power capacity equal to 9.2 kW and an offset equal to 0.8 kW. From (3) the droopis derived and shown in Figure 4.
In this paper, the Danish frequencies, both DK1 and DK2 [23] (DK1 is the western Denmark partof the European power system, and DK2 is the eastern Denmark part of the Scandinavian countries.The two areas are electrically connected through the DC Great Belt Power Line), and the Japanese(JP) ones are considered. Figure 5 shows the 13 h of frequency both as Gaussian distribution and asfrequency versus time.
World Electric Vehicle Journal 2020, 11, 48 6 of 15
49.8 49.9 50 50.1 50.2−12−10−8−6−4−2
02468
10
Frequency [Hz]
P bat
t[k
W]
Figure 4. Droop characteristic.
49.8 49.9 50 50.1 50.20
5,000
10,000
15,000
20,000Std: 0.0175Mean: 50.0001
Frequency [Hz]
Cou
nts
DK1
0 4 8 12 16 20 2449.8
49.9
50
50.1
50.2
Time [h]
Freq
uenc
y[H
z]
49.8 49.9 50 50.1 50.20
5,000
10,000
15,000
20,000
25,000 Std: 0.043Mean: 50.0037
Frequency [Hz]
DK2
0 4 8 12 16 20 2449.8
49.9
50
50.1
50.2
Time [h]
49.8 49.9 50 50.1 50.20
5,000
10,000
15,000
20,000
25,000 Std: 0.0261Mean: 50.0007
Frequency [Hz]
JP
0 4 8 12 16 20 2449.8
49.9
50
50.1
50.2
Time [h]
Figure 5. Frequency histogram fitted in Gaussian distribution and frequency trend during the 13 h ofFR service provision in DK1, DK2 and Japan.
Given the frequency, the battery and grid powers are derived as in (3) and (4), considering theefficiency of the ±10 kW bidirectional charger as in Table 1. The battery power (Pbatt) for one-dayperiod of the driv scenario and the driv + FR scenario for the three areas DK1, DK2 and JP are providedin Figure 6.
When providing FR, even though the battery could be charged from 6:00 to 8:00, in all three casesthe charging time is very short in comparison to the driving scenario. This is due to the fact that thebattery is charged more than what it is discharged during the FR hours, thus there is less need forcharging to the required SOC level. This is in accordance with the frequency mean values provided inFigure 5, which are slightly higher than 50 Hz in all the three areas.
Table 1. Bidirectional 10 kW DC charger efficiency [8].
Pgch 0 −0.78 −0.91 −1.15 −1.49 −1.99 −2.63 −3.67 −4.68 −5.94 −6.9 −7.97 −9.09 −10ηch 0.01 0.11 0.27 0.42 0.54 0.66 0.73 0.79 0.82 0.85 0.86 0.87 0.87 0.87
Pgdisch 0 0.17 0.33 0.59 0.92 1.29 1.87 2.42 3.27 4.25 5.27 6.08 7.49 8.71ηdisch 0.01 0.14 0.3 0.45 0.57 0.67 0.74 0.78 0.82 0.84 0.85 0.86 0.86 0.88
World Electric Vehicle Journal 2020, 11, 48 7 of 15
Figure 6. Battery power of the three scenarios during one-day period: driving, driving + FR in DK1,DK2 and JP.
3.1.2. Input: Temperature
Figure 7 shows the outside temperature during year 2017 in Denmark, which is considered duringthe simulations for DK1, DK2 and JP.
Jan FebMar
AprMay Jun Jul
Aug Sep OctNov
Dec−20−10
0102030
T out
[o C]
Figure 7. Outside temperature year 2017 in Denmark.
3.2. Economic Characterization
3.2.1. Frequency Regulation Price
The FR market framework in DK2 is taken as reference for all cases, due to the fact that JP doesnot have a market yet, and in DK1 the prices over the last 10 years are similar to the DK2 ones [24].The average hourly price of frequency containment reserves in DK1 from 2015 to 2019 is providedin Figure 8. In 2015 the prices were exceptionally low due to high rainfall in the Nordic region,where hydro plants lead the provision of FR decreasing the prices [8,25]. By contrast, 2018 was a “dry”year, where the less FR from hydro plants increased the FR provided by the conventional units, and sodid the prices.
World Electric Vehicle Journal 2020, 11, 48 8 of 15
0 2 4 6 8 10 12 14 16 18 20 22 240
20
40
60
Time [h]
Pric
e[¤
/MW
] 2015 20162017 20182019
Figure 8. Average hourly prices of FR in DK2 [24,26].
3.2.2. Electricity Price
In 2018–2019 the electricity price for the Danish domestic consumers was 0.31 ¤/kWh,included taxes, whereas the industrial one (mid-size industry) was 0.08 ¤/kWh, excluded taxesas they are not applied to industrial prices. For what concerns the sale electricity price, over the last 10years the average price was approx. 0.04 ¤/kWh.
3.2.3. Battery Price
For the evaluation of the BD costs, the value of 180 ¤/kWh is considered to be battery pack priceper kWh [27]. Nevertheless, this is a conservative assumption since batteries are becoming cheaperand the price is expected to drop to less than 100 ¤/kWh by 2030 [28,29].
4. Results
During the following analysis, the battery model is considered to receive as input the same outsidetemperature and battery power for the 5 years of simulation. The scenario driv is run with both SOCequal to 55 and 75%, and it is the same for all the three areas. The scenario driv + FR is run for all thethree areas—DK1, DK2 and JP—with both SOC equal to 55 and 75%.
4.1. Technical Results
Figure 9 provides the SOH of the battery over the five years in the different scenarios:
1 2 3 4 5
90
95
100
SOH
[%]
Mean SOC ~ 55%
Driv Driv+FRDK1Driv+FRDK2 Driv+FRJP
1 2 3 4 5
90
95
100
Year
SOH
[%]
Mean SOC ~ 75%
Figure 9. SOH comparison of the driv and driv + FR scenarios over five years in DK1, DK2 and JP:subplot 1 with average SOC 55% and subplot 2 with average SOC 75%.
World Electric Vehicle Journal 2020, 11, 48 9 of 15
4.2. Economic Results
In this section, the profitability of the FR service is quantified. Revenue, costs and profits aredetermined as described in Section 2. Table 2 provides the RevenueFR, which is independent on theSOC, electricity price and area.
Table 2. RevenueFR for the 5 years of frequency regulation and sum of the 5 years.
Year 2015 2016 2017 2018 2019 meanRevenueFR [¤/y] 667 1324 1226 1925 1466 1322
RevenueFR [¤/5 y] 6608
Table 3 shows the costs due to energy exchanged with the grid and charger loss, comparing thecase with domestic and industrial electricity prices and the three considered areas. The domesticand industrial electricity prices in DK2 are considered for all three areas. Since the power profile ofthe battery is the same for the 5 years, the cost for 5 years is the CostFR
loss for one year multiplied by5. It is relevant to observe that with domestic electricity prices the CostFR
loss in DK2 is higher than inDK1 and JP, which is not the case when the industrial electricity prices are considered. This is due totwo main reasons: first the frequency in DK2 has a standard deviation of 0.043, larger than in DK1and JP, meaning that the power request during the regulation is larger. With larger power, both forcharging and discharging, losses are lower, because the charger efficiency is closer to the optimal.Second, the domestic electricity price during purchase is higher than the electricity price for sale,making the purchasing component large in comparison to the selling one. The combination of thesetwo shows that during FR the purchase/sale of electricity is causing costs, which are much larger inthe domestic case.
Table 3. CostFRloss for 5-year period of frequency regulation, due to the charging and discharging
charger loss.
Electricity Price Domestic Industrial
Country DK1 DK2 JP DK1 DK2 JPCostFR
loss [¤/5 y] 12,065 13,260 12,815 2970 2860 3000
The difference between the battery power in the three areas is graphically provided in Figure 10as battery power in function of the gradient of the battery power.
Figure 10. Battery power as function of the gradient of the battery power for DK1, DK2 and JP.
Table 4 provides the costs due to battery degradation when just driving, driving plus FR and justFR. The first two costs are given to underline that the degradation is a relevant cost during the lifetimeof the battery, nevertheless the FR provision is not adding a large component to the driv scenario.Furthermore, a difference can be observed between DK2 and the other two areas: in DK2 CostFR
BDrepresents approximately 16–20% of the total Costdriv+FR
BD , depending on the SOC level. Differently
World Electric Vehicle Journal 2020, 11, 48 10 of 15
in DK1 and JP the FR represents approx. the 10% of the total costs. It is important to highlight that,in contrast to the CostFR
loss, the costs due to BD are not the same for all the 5 years. This is due to thecalendar and cycling degradation, which have an exponential behavior during the first years of thebattery lifetime, see Figure 9.
Table 4. CostdrivBD , Costdriv+FR
BD and CostFRBD for 5 years period of battery usage.
Mean SOC 55% 75%
Country DK1 DK2 JP DK1 DK2 JPCostdriv
BD [¤/5 y] 876 1314Costdriv+FR
BD [¤/5 y] 968 1081 1019 1418 1565 1438CostFR
BD [¤/5 y] 92 205 143 104 251 124
Figure 11 summarizes in the first subplot revenues and costs of the different cases whereas thesecond subplot compares the profits during the five years period, as calculated with (1). The differencebetween the SOC, with both domestic and industrial electricity prices, does not have a large impacton the final profitability of the service. By contrast, the energy exchanged with the grid is substantial,causing economic loss when the domestic prices are considered. For what concerns the industrialelectricity prices, the cost of the purchase is one fourth of the domestic one and thus the service isprofitable with approx. 3500¤ for the considered five years and in the different scenarios.
Figure 11. Comparison of domestic and industrial cases, in black for SOC equal to 55% and in redfor SOC equal to 75%, in DK1, DK2 and JP; revenues and costs in the first subplot, profits in thesecond subplot.
5. Sensitivity Analysis
In this section, three parameters are further investigated, as they provide further highlights forthe evaluation of the BD losses and relative costs.
World Electric Vehicle Journal 2020, 11, 48 11 of 15
5.1. Outside Temperature
The outside temperature is a relevant parameter during the calculation of the battery degradation,because it influences both the calendar and the cycle processes, as shown in (6) and (7). In this regards,the authors selected the JP case with SOC equal to 55% and investigated the battery degradation whenthe temperature is increased of 5, 10 and 15 ◦C throughout the entire year. Figure 12 compares theresults of the different cases, first for the calendar and the cycle degradation and then for the SOH.Results show that despite the temperature increase of 5, 10 and 15 ◦C, the maximum SOH differenceis 3%, which would result on a contained increase of the BD costs. Furthermore, the calendar lossesare observed to be predominant in the total loss and their increase goes with the increase of thetemperature. By contrast, the cycle degradation for the 10 and 15 ◦C temperature increase is almost thesame. This is explained by the relation between the cycle degradation and the temperature providedin Figure 2, which shows that around 25 ◦C there is a minimum, and then the degradation starts toincrease again.
1 2 3 4 502468
10
Qca
l[%
]
Base case + 5oC + 10oC + 15oC
1 2 3 4 50
0.51
1.52
2.5
Qcy
cle
[%]
1 2 3 4 590
95
100
Year
SOH
[%]
Figure 12. Comparison between the JP base case with SOC equal to 55% and the cases with outsidetemperature increase of 5, 10 and 15 ◦C: subplot 1 calendar degradation, subplot 2 cycle degradationand subplot 3 SOH.
5.2. Used Capacity
The second parameter is the remaining capacity of the battery. As shown in Figure 3, the batteryend-of-life is set to SOH of 50%, after the battery has been used in the vehicle and for a second use.Nevertheless, as presently, not all batteries used in EVs are then used in second life applications.To consider this case the 50% remaining battery capacity in (12) is compared with the 20% usedcapacity, when the battery end-of-life coincides with the EV end-of-life (final SOH = 80%). The CostFR
BDof the initial case (CostFR
BD50%) is compared with the CostFR
BD20%in Table 5:
World Electric Vehicle Journal 2020, 11, 48 12 of 15
Table 5. CostFRBD for 5 years period of battery usage considering the used capacity equal to 20%, 50%.
Mean SOC 55% 75%
Country DK1 DK2 JP DK1 DK2 JP
CostFRBD20%
[¤/5 y] 230 513 358 260 628 310CostFR
BD50%[¤/5 y] 92 205 143 104 251 124
The results show that the battery degradation is max 5% of the total cost when consideringthe domestic charger loss present in Table 3. When considering the industrial prices, the batterydegradation could represent up to 20% of the cost, resulting on lower profit, but still positive.
5.3. Battery Price
To evaluate the BD costs, the battery price per kWh has been considered equal to 180¤. However,the price of NMC batteries is decreasing year-by-year and forecasts show that the price could becomelower than 90 ¤/kWh by 2030 [29]. Furthermore, to account for price variations among automakersand the price difference between the cost for the automaker and the costumer, prices higher than180 ¤/kWh are also considered. In this regard, Figure 13 compares the CostFR
BD evaluated in Table 4,with the BD cost for FR when the battery price per kWh is equal to 240, 210, 180, 150, 120 and 90¤.The BD cost due to FR are shown to decrease to less than 100¤ for DK1 and JP, meaning that in thefuture their weight on the final profit can become even lower than what shown in Figure 11.
240 210 180 150 120 900
100
200
300
400
Battery Price [¤/kWh]
Cos
tFR BD[¤
]
DK1DK2JP
Figure 13. Comparison of CostFRBD with battery prices equal to 240, 210, 180, 150, 120 and 90 ¤/kWh.
In black the cases with SOC = 75% and in red the ones with SOC = 55%.
6. Conclusions
In this manuscript the profitability of primary frequency regulation provided by EVs, taking intoaccount the battery degradation costs has been quantified based on the frequency measured inDenmark, both DK1 and DK2, and Japan. First, the battery power profiles have been derived from thedriving pattern and the frequency measurements. Second, the battery model has been simulated toderive the degradation loss and to calculate the related costs the user would encounter while providingfrequency regulation. Afterwards the energy exchanged with the grid has been used to determinethe economic loss due to the higher purchase electricity price in comparison to the sale one. Finally,revenue and costs were combined to determine the profit. It was noticed that the frequency in DK2was more spread between the 49.8 and 50.2 Hz causing higher costs due to degradation and energyexchanged with the grid. By contrast, for what concerns the battery degradation, the losses are largerwhen the frequency is more spread, because the battery is exposed to more cycles during its lifetime.Nevertheless, the battery degradation cost is approx. 1% of the total costs when considering domesticelectricity prices, and 3–8% when considering industrial prices, both with 55 and 75% SOC.
World Electric Vehicle Journal 2020, 11, 48 13 of 15
Finally, it was observed that the electricity price and the charger efficiency are two main parametersthat affect the profitability of the service. Indeed, with domestic prices the charger losses are too largeand there are only economic losses in the service provision.
Further considerations and analysis can be done regarding the losses:
• charger efficiency: the charger losses of this analysis are based on the efficiency of a three yearsold charger. It is interesting to compare the charger efficiency of this charger with new and futurechargers, which could have higher efficiency and thus lower costs [30].
• battery capacity: the considered battery has a 40 kWh capacity. Considering the samecharging/discharging power profile, larger is the battery, lower is the amount of cycles thebattery is exposed to. Thus, for smaller batteries the battery degradation is expected to be a largercomponent of the total amount of loss. For larger batteries, which are expected in the future,the degradation can be even lower, decreasing further the degradation costs. This would also givemore flexibility on the charging power that can be used during the frequency regulation provision.
Even though DK1 (belonging to the continental Europe synchronous areas) has been consideredduring the analyses, to generalize the results to the all Europe the following aspects should becarefully considered:
• Cost of electricity: in Europe it ranges around 0.21 €/kWh whereas in Denmark it is approx.0.31 €/kWh, resulting on lower cost of electricity in most European countries.
• Vehicle driving pattern and annual distance: European countries have similar driving patterns,nevertheless distances ranges between 40 and 80 km/day depending on the consideredcountry [31].
• Ambient temperature: European countries are characterized by wide variety of climate zones,meaning that the temperature behavior throughout the year can vary greatly.
Similar aspects must be considered when generalizing the analyses of DK2 to the Scandinaviancountries. Further studies of the authors will include the different considerations for the evaluation ofthe profitability of frequency regulation.
Author Contributions: Conceptualization, L.C. and M.M.; data curation, L.C.; investigation, L.C. and M.M.;methodology, L.C. and M.M.; writing—original draft preparation, L.C.; writing—review and editing, L.C. andM.M. All authors have read and agreed to the published version of the manuscript.
Funding: The work in this paper has been supported by the research projects ACES (EUDP grant nr:EUDP17-I-12499) and CAR (EU-Interreg grant nr: STHB.03.01.00-SE-S112/17).
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Kempton, W.; Tomic, J. Vehicle-to-grid power fundamentals: Calculating capacity and net revenue.J. Power Sources 2005, 144, 268–279. [CrossRef]
2. Andersen, P.B.; Sousa, T.; Thingvad, A.; Berthou, L.S.; Kulahci, M. Added Value of Individual FlexibilityProfiles of Electric Vehicle Users for Ancillary Services. In Proceedings of the 2018 IEEE InternationalConference on Communications, Control, and Computing Technologies for Smart Grids, SmartGridComm2018, Aalborg, Denmark, 29–31 October 2018; [CrossRef]
3. Thompson, A.W. Economic implications of lithium ion battery degradation for Vehicle-to- Grid (V2X)services. J. Power Sources 2018, 396, 691–709. [CrossRef]
4. Thompson, A.W.; Perez, Y. Vehicle-to-Everything (V2X) energy services, value streams, and regulatorypolicy implications. Energy Policy 2020, 137, 111136. [CrossRef]
5. Thingvad, A.; Martinenas, S.; Andersen, P.B.; Marinelli, M.; Olesen, O.J.; Christensen, B.E. Economiccomparison of electric vehicles performing unidirectional and bidirectional frequency control in Denmarkwith practical validation. In Proceedings of the 51st International Universities Power Engineering Conference,UPEC 2016, Coimbra, Portugal, 6–9 September 2017. [CrossRef]
World Electric Vehicle Journal 2020, 11, 48 14 of 15
6. Codani, P.; Perez, Y.; Petit, M. Financial shortfall for electric vehicles: Economic impacts of TransmissionSystem Operators market designs. Energy 2016, 113, 422–431. [CrossRef]
7. Zecchino, A.; Thingvad, A.; Andersen, P.B.; Marinelli, M. Test and modelling of commercial V2G CHAdeMOchargers to assess the suitability for grid services. World Electr. Veh. J. 2019, 10. [CrossRef]
8. Thingvad, A.; Ziras, C.; Marinelli, M. Economic value of electric vehicle reserve provision in the Nordiccountries under driving requirements and charger losses. J. Energy Storage 2019, 21, 826–834. [CrossRef]
9. de la Prieta, F.; Escalona, M.J.; Corchuelo, R.; Mathieu, P.; Vale, Z.; Campbell, A.T.; Rossi, S.; Adam, E.;Jiménez-López, M.D.; Navarro, E.M.; et al. Trends in Practical Applications of Scalable Multi-Agent Systems,the PAAMS Collection; Springer International Publishing: Cham, Switzerland, 2016; Volume 473. [CrossRef]
10. Thingvad, A.; Calearo, L.; Andersen, P.B.; Marinelli, M.; Neaimeh, M.; Suzuki, K.; Murai, K. Valueof V2G Frequency Regulation in Great Britain Considering Real Driving Data. In Proceedings of the2019 IEEE PES Innovative Smart Grid Technologies Europe, ISGT-Europe 2019, Bucharest, Romania, 29September–2 October 2019; pp. 1–5, [CrossRef]
11. Petit, M.; Perez, Y. Vehicle-to-grid in France: What revenues for participation in frequency control?In Proceedings of the International Conference on the European Energy Market, EEM, Stockholm, Sweden,27–31 May 2013; pp. 1–7, [CrossRef]
12. Dallinger, D.; Krampe, D.; Wietschel, M. Vehicle-to-grid regulation reserves based on a dynamic simulationof mobility behavior. IEEE Trans. Smart Grid 2011, 2, 302–313. [CrossRef]
13. Smith, K.; Earleywine, M.; Wood, E.; Neubauer, J.; Pesaran, A. Comparison of plug-in hybrid electric vehiclebattery life across geographies and drive cycles. SAE Tech. Pap. 2012. [CrossRef]
14. Hicks-Garner, J.; Wang, J.; Soukazian, S.; Purewal, J.; Tataria, H.; Liu, P.; Sorenson, A.; Sherman, E.;Verbrugge, M.W.; Vu, L. Degradation of lithium ion batteries employing graphite negatives andnickel–cobalt–manganese oxide+spinel manganese oxide positives: Part 1, aging mechanisms and lifeestimation. J. Power Sources 2014, 269, 937–948. [CrossRef]
15. Mathieu, R.; Baghdadi, I.; Briat, O.; Gyan, P.; Vinassa, J.M. D-optimal design of experiments applied tolithium battery for ageing model calibration. Energy 2017, 141, 2108–2119. [CrossRef]
16. Calearo, L.; Thingvad, A.; Marinelli, M. Modeling of Battery Electric Vehicles for Degradation Studies.In Proceedings of the 54th International Universities Power Engineering Conference, UPEC 2019, Bucharest,Romania, 3–6 September 2019. [CrossRef]
17. Thingvad, A.; Marinelli, M. Influence of V2G Frequency Services and Driving on Electric Vehicles BatteryDegradation in the Nordic Countries Influence of V2G Frequency Services and Driving on Electric VehiclesBattery Degradation in the Nordic Countries. In Proceedings of the 31st International Electric VehiclesSymposium & Exhibition & International Electric Vehicle Technology Conference 2018, Kobe ConventionCenter, Kobe, Japan, 30 September 2018–3 October 2018.
18. Thingvad, M.; Calearo, L.; Thingvad, A.; Marinelli, M. Characterization of NMC Lithium-ion BatteryDegradation for Improved Online Estimation. In Proceedings of the 55th International Universities PowerEngineering Conference, UPEC2020, Torino, Italy, 1–4 September 2020; in press.
19. ACES Project Website. Available online: https://www.aces-bornholm.eu/ (accessed on 1 July 2020).20. Wang, D.; Coignard, J.; Zeng, T.; Zhang, C.; Saxena, S. Quantifying electric vehicle battery degradation from
driving vs. vehicle-to-grid services. J. Power Sources 2016, 332, 193–203. [CrossRef]21. Calearo, L.; Thingvad, A.; Marinelli, M. Experimental Validation of Lithium-ion Battery Model for
Degradation Studies with no Direct Access to the Battery Pack. Unpublished work, 2020.22. Calearo, L.; Thingvad, A.; Suzuki, K.; Marinelli, M. Grid Loading Due to EV Charging Profiles Based on
Pseudo-Real Driving Pattern and User Behavior. IEEE Trans. Transp. Electrif. 2019, 5, 683–694. [CrossRef]23. Thingvad, A.; Marinelli, M. Grid Frequency Measurements of the Nordic Power System during 2018. Dataset,
5 May 2020. [CrossRef]24. Frequency Containment Reserves (FCR). DK2. Available online: https://www.energidataservice.dk/
(accessed on 3 July 2019).25. Ropenus, S.; Jacobsen, H.K. A Snapshot of the Danish Energy Transition: Objectives, Markets, Grid, Support
Schemes and Acceptance Study; Agora Energiewende: Berlin, Germany and DTU Management Engineering:Lyngby, Denmark, 2015.
26. Electricity Price Statistics. Available online: https://ec.europa.eu/eurostat/statistics-explained/ (accessedon 3 June 2019).
World Electric Vehicle Journal 2020, 11, 48 15 of 15
27. Wentker, M.; Greenwood, M.; Leker, J. A bottom-up approach to lithium-ion battery cost modeling with afocus on cathode active materials. Energies 2019, 12, 504. [CrossRef]
28. Berckmans, G.; Messagie, M.; Smekens, J.; Omar, N.; Vanhaverbeke, L.; Mierlo, J.V. Cost projection of state ofthe art lithium-ion batteries for electric vehicles up to 2030. Energies 2017, 10. [CrossRef]
29. Tsiropoulos, I.; Tarvydas, D.; Lebedeva, N. Li-Ion Batteries for Mobility and Stationary StorageApplications—Scenarios for Costs and Market Growth; EUR 29440 EN; Publications Office of the EuropeanUnion: Luxembourg, 2018; ISBN 978-92-79-97254-6. [CrossRef]
30. Genovese, A.; Ortenzi, F.; Villante, C. On the energy efficiency of quick DC vehicle battery charging.World Electr. Veh. J. 2015, 7, 570–576. [CrossRef]
31. Pasaoglu, G.; Fiorello, D.; Martino, A.; Scarcella, G.; Alemanno, A.; Zubaryeva, A.; Thiel, C. Driving andParking Patterns of European Car Drivers—A Mobility Survey; Publications Office of the European Union:Luxembourg, 2012; p. 112, [CrossRef]
© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open accessarticle distributed under the terms and conditions of the Creative Commons Attribution(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
